You are building a fraud detection model using Snowflake data'. The dataset 'TRANSACTIONS' contains billions of records and is partitioned by 'TRANSACTION DATE'. You want to use cross-validation to evaluate your model's performance on different subsets of the data and ensure temporal separation of training and validation sets. Given the following Snowflake table structure:

Which approach would be MOST appropriate for implementing time-based cross-validation within Snowflake to avoid data leakage and ensure robust model evaluation? (Assume using Snowpark Python to develop)

• A. Explicitly define training and validation sets based on date ranges within the Snowpark Python environment, performing iterative training and evaluation within the client environment before deploying a model to Snowflake. No built-in cross-validation used
• B. Create a UDF that assigns each row to a fold based on the 'TRANSACTION DATE column using a modulo operation. This is then passed to the 'cross_validation' function in Snowpark ML.
• C. Utilize the 'SNOWFLAKE.ML.MODEL REGISTRY.CREATE MODEL' with the 'input_colS argument containing 'TRANSACTION DATE'. Snowflake will automatically infer the temporal nature of the data and perform time-based cross-validation.
• D. Use 'SNOWFLAKE.ML.MODEL REGISTRY.CREATE MODEL' with default settings, which automatically handles temporal partitioning based on the insertion timestamp of the data.
• E. Implement a custom splitting function within Snowpark, creating sequential folds based on the 'TRANSACTION DATE column and use that with Snowpark ML's cross_validation. Ensure each fold represents a distinct time window without overlap.

You have deployed a fraud detection model in Snowflake and are monitoring its performance. The initial AUC was 0.92. After a month, you observe the AUC has dropped to 0.78. You suspect data drift. Which of the following steps should you take FIRST to investigate and address this performance degradation, focusing on efficient resource utilization within Snowflake?

• A. Increase the complexity of the existing model architecture by adding more layers to the neural network to improve its adaptability.
• B. Analyze the distributions of key features in the current production data compared to the training data using Snowflake SQL queries and visualization tools. Specifically compare the distributions of features such as transaction amount and time of day. Then, if drift is confirmed, retrain using updated data.
• C. Delete the existing model and deploy a pre-trained, generic fraud detection model obtained from a public repository.
• D. Deploy a new model version with a higher classification threshold to compensate for the increased false positives.
• E. Immediately retrain the model using the entire dataset available, scheduling a Snowpark Python UDF to perform the training.

You are tasked with performing exploratory data analysis on a table named containing daily sales transactions. The table includes columns like 'transaction_date', 'product_id', 'quantity' , and 'price'. Your goal is to identify potential data quality issues and understand the distribution of sales. Which of the following SQL queries using Snowflake's statistical functions and features would be MOST effective for quickly identifying outliers in the 'quantity' column, potential data skewness, and missing values?

• A. Option B
• B. Option E
• C. Option D
• D. Option C
• E. Option A

You are working with a large dataset of transaction data in Snowflake to identify fraudulent transactions. The dataset contains millions of rows and includes features like transaction amount, location, time, and user ID. You want to use Snowpark and SQL to identify potential outliers in the 'transaction amount' feature. Given the potential for skewed data and varying transaction volumes across different locations, which of the following data profiling and feature engineering techniques would be the MOST effective at identifying outlier transaction amounts while considering the data distribution and location-specific variations?

• A. Apply a clustering algorithm (e.g., DBSCAN) using Snowpark ML to the transaction data, using transaction amount, location and time as features. Treat data points in small, sparse clusters as outliers. This approach does not need to be performed for each location, just the entire dataset.
• B. Partition the data by location using Snowpark. For each location, calculate the median and median absolute deviation (MAD) of the 'transaction amount' feature. Identify outliers as transactions with amounts that fall outside of the median +/- 3 MAD for that location.
• C. Use Snowflake's APPROX_PERCENTILE function with Snowpark to calculate percentiles of the 'transaction amount' feature. Transactions with amounts in the top and bottom 1% are flagged as outliers.
• D. Use Snowpark to calculate the interquartile range (IQR) of the 'transaction amount' feature for the entire dataset. Identify outliers as transactions with amounts that fall below QI - 1.5 IQR or above Q3 + 1.5 IQR.
• E. Calculate the mean and standard deviation of the 'transaction amount' feature for the entire dataset using SQL. Identify outliers as transactions with amounts that fall outside of 3 standard deviations from the mean.

You're building a regression model using Snowpark Python to predict house prices. After initial training, you observe that the model consistently overestimates the prices of high-value houses and underestimates the prices of low-value houses. Given the options below, which optimization metric, along with code snippet to calculate it using Snowpark, would be most effective in addressing this specific issue?

• A. Root Mean Squared Error (RMSE) - as it gives more weight to larger errors, making it suitable for addressing the underestimation/overestimation problem.
  
• B. Adjusted R-squared - as it penalizes the addition of irrelevant features, improving the model's generalization ability.
  
• C. R-squared - as it measures the proportion of variance explained, directly addressing how well the model fits the data across all price ranges.
  
• D. Mean Squared Error (MSE) - as it is less sensitive to outliers than RMSE.
  
• E. Mean Absolute Error MAE - as it is sensitive to outliers and will penalize large errors more heavily.